The SpaceX IPO:
What We’re Watching

SpaceX filed its S-1 on May 20 and is expected to begin trading on the Nasdaq under the ticker SPCX in mid-June. At a reported valuation of $1.75 to $2 trillion and a raise that could approach $75 billion, this would be the biggest IPO ever — several times the size of the previous record holder, Saudi Aramco. A deal this large, this hyped, and this concentrated under one founder deserves a clear-eyed look. Here’s what the filing actually shows.

The dates and deadlines that matter

The process has moved faster than most expected: confidential filing on April 1, public prospectus on May 20, and a targeted debut just over three weeks later.

Two of these matter more than the splashy debut. The November earnings report is the first time the market grades the company on audited numbers rather than a prospectus narrative, and the December lockup expiration is when insider supply unlocks. Day one is theater; these are where price discovery actually happens.

Indications of conditional interest: June 4–10

This IPO is unusual in how openly it courts individuals. SpaceX stated in the filing that a portion of Class A shares will be offered to retail investors, routed through major brokerage platforms — reportedly around 30% of the deal, roughly triple the historical norm.

A few caveats. Eligibility and minimums vary by platform — reports point to asset thresholds at some brokerages (Schwab around $100K in eligible assets, Fidelity reportedly higher) and suitability questionnaires at others. Demand will far exceed supply, so a request is not a fill: most retail investors should expect a partial allocation, or none. And submitting an indication is not a commitment to a price you haven’t seen — the range is set during the window, and you affirm afterward. Your advisor can typically submit the request on your behalf.

What the price tag rests on

The S-1 cover page is deliberately blank on price — standard practice. The actual range gets set during the roadshow based on demand, with pricing expected around June 11. The scale is the story: this single deal could raise more than the two largest IPOs in history combined.

History is also a warning. First-day pops on hot debuts have averaged roughly 19% since 1980, but high-profile names frequently retrace 20%+ within their first few months. Be cautious in sizing any IPO position and expect significant volatility.

Three businesses sold as one number

The $1.75–2 trillion figure rests on about $18.7 billion of 2025 revenue — a number that now folds in xAI and X — against a GAAP net loss near $4.9 billion. Underneath the headline are three very different businesses.

The read is straightforward: you’re buying a world-class satellite-internet business at a venture-stage multiple, and the trillion-dollar premium on top is a bet on Starship, orbital compute, and AI paying off years from now.

An easier way to see how steep that is: price-to-sales — simply what you pay for every dollar of revenue. At a ~$1.75–2 trillion valuation on about $18.7 billion of 2025 revenue, SpaceX would come public at roughly 100 times sales. For comparison, the S&P 500 trades around 3.7× sales, Apple around 10×, and even Nvidia — the poster child of the AI boom — sits near 24×. SpaceX would debut at four times Nvidia’s multiple and well above almost anything in public markets. A number that high only makes sense if you believe revenue grows many times over from here; you’re paying today for a business that mostly doesn’t exist yet.

The other side: why someone pays 100× sales

That multiple isn’t indefensible — it’s a bet that today’s revenue is a rounding error against what’s coming. The bull case rests on a simple idea: SpaceX isn’t valued as a company, it’s valued as infrastructure for several trillion-dollar markets at once, most of which it already dominates. Walk through the drivers and the number starts to look less crazy.

Starlink is compounding, not maturing. The subscriber base crossed 10 million in early 2026 and has roughly doubled annually. Independent projections see 18–20 million subscribers in 2026 and as many as 60 million by 2030 — revenue scenarios that run from $40 billion to $100 billion a year from connectivity alone. Morgan Stanley models Starlink at ~72% of a $65 billion SpaceX by 2030; ARK’s base case puts the whole company at $2.5 trillion by then. At $65–100 billion of revenue, today’s ~100× multiple compresses to single digits without the stock moving at all. That’s the entire bull thesis in one sentence: grow into the number.

The new revenue lines barely exist yet. Direct-to-cell — satellite service straight to an ordinary phone, no dish — turns every smartphone on Earth into a potential customer. SpaceX bought $17 billion of EchoStar spectrum in 2025 to build it, with T-Mobile already partnered; estimates for the opportunity run into the tens of billions annually by 2030. Add high-margin government and defense work (Starshield), maritime and aviation connectivity, and you have multiple subscription businesses layered on the same constellation, each at telecom-utility margins. Starlink already runs a ~60%+ EBITDA margin — rare for anything growing this fast.

Starship is the asymmetric bet. If reusable Starship works at scale, the cost to put mass in orbit collapses — and everything downstream (bigger Starlink satellites, orbital data centers, lunar and Mars logistics, launch for everyone else) gets cheaper and bigger at the same time. SpaceX already owns the launch market; Starship widens that moat into businesses that don’t have names yet. The S-1 frames the combined SpaceX–xAI opportunity as a $26–28 trillion total addressable market, with on-orbit AI compute — “orbital intelligence” — as the connective tissue. Take even a sliver of that and the math works.

None of this is guaranteed, and the section above lays out why the price leaves no room for error. But the honest framing is this: you’re not paying 100× for what SpaceX is — you’re paying it for a credible path to being a $300 billion-revenue company that owns the plumbing of global connectivity, launch, and orbital compute. Whether that path is worth the price is the question. That it’s a plausible path is not really in dispute.

Profitable engine, loss-making whole

Whether SpaceX “makes money” depends on which line you read. On an adjusted-EBITDA basis it is solidly positive; on a GAAP basis — after stock comp, constellation depreciation, and AI capex — it posted a sizeable loss. The xAI merger turned a profitable launch-and-connectivity company into a heavy loss-maker overnight, by design.

The AI segment is the primary loss driver, burning in the $6B+ range annually — even as it carries large contracted revenue, including a reported ~$1.25B/month data-center deal running through 2029. Big numbers on both sides of the ledger.

The empire and the Tesla question

The IPO may be a step toward something larger, not the end state. SpaceX already absorbed xAI — which had absorbed X — consolidating rocket, satellite, social, and AI assets under one roof. The chatter now is the next consolidation: folding Tesla into the structure.

The logic is real. The two companies share engineers, board members, and a joint chip-fabrication venture in Texas. Taking SpaceX public at a valuation in Tesla’s neighborhood makes a stock-for-stock combination far easier to contemplate — and a merger would hand Musk the super-voting control he holds at SpaceX but lacks at Tesla. Early backer Peter Diamandis has framed it as a matter of when, not if; Wedbush’s Dan Ives expects the two to combine within about a year. Nothing is confirmed, and the conflict-of-interest questions would be significant — but it’s a scenario worth tracking for anyone holding Tesla.

One thing not up for debate: control

This is a controlled company. Public Class A shares carry one vote each; Musk’s Class B shares carry ten. Buying SPCX is a bet on Musk’s vision — you don’t get a meaningful say in how it’s executed.

What that actually means for a shareholder

In practice, super-voting control means the things shareholders normally get a say in — board composition, executive pay, major acquisitions, even removing a CEO who is underperforming — are effectively decided by the founder regardless of how public holders vote. If you disagree with a strategic direction (say, folding Tesla in, or pouring billions more into xAI), your recourse is to sell, not to vote. You are, quite literally, along for the ride. The flip side: you also can’t be steamrolled by an activist or a short-term board pushing the company off its long-term plan. Whether that’s reassuring or alarming depends entirely on how much you trust the person at the wheel.

Hasn’t this played out fine before?

Often, yes — which is exactly why founders keep doing it. This structure isn’t exotic. Google pioneered the modern version at its 2004 IPO, giving Larry Page and Sergey Brin Class B shares with 10 votes each (and later adding a third, no-vote class); Meta runs the same playbook, with Mark Zuckerberg’s Class B stock holding roughly 10 votes per share and majority control despite a minority economic stake. Alphabet, Meta, Palantir, and Berkshire Hathaway all carry multiple share classes today. For long-term holders of Google and Meta, ceding voting power has been largely irrelevant next to the returns — and that track record is precisely the bull case for trusting a proven founder.

The broader data leans the same way, with caveats. A 2024 analysis in the Harvard Law School Forum on Corporate Governance found that, on average, Russell 3000 companies with dual- or multi-class structures outperformed single-class peers over five- and ten-year windows. But a separate European Corporate Governance Institute study found the valuation premium of dual-class firms tends to fade over time — the founder’s edge is most valuable early, and the same insulation that protects a visionary can later entrench a management team that has stopped adapting. Governance watchdogs like the Council of Institutional Investors argue that the “founder-knows-best” model can blind a company to the need for change.

So the honest answer is: it has worked well when the founder kept delivering, and poorly when they didn’t — and shareholders had no lever to pull either way.